Unprecedented large-area turnover estimates for the subtropical Brazilian Atlantic Forest based on systematically-gathered data

Abstract

Secondary forests play an important role as landscape elements, providing carbon uptake, biodiversity recovery, and income for rural communities. We aimed to present change estimates for three forest types in the subtropical Atlantic Forest, namely evergreen rainforest (ERF), Araucaria forest (AF), and semi-deciduous forest (SF). We used data gathered on 303 systematically-distributed permanent plots distributed across ∼23% of the subtropical Brazilian Atlantic Forest, which nowadays is mostly composed of secondary stands. We found positive mean net changes in tree density ha−1 yr−1, basal area ha−1 yr−1, and aboveground biomass (AGB) ha−1 yr−1 for all forest types. The ERF had a statistically significantly greater mean net change in AGB ha−1 yr−1 than those of the other forest types. Relative to initial and final stocks though, mean annual rates of gain and loss (% ha−1 yr−1) for the abovementioned forest attributes were statistically equivalent among the forest types. Estimated mean net changes in AGB for the studied forest types ranged from 4.9 to 8.1 Mg ha−1 for a 5-year period. The mean net changes we found are smaller than those reported for secondary tropical forests, possibly due to climate constraints and increased sampling variability. The AGB stocks of the inventoried forests accounted, on average, to one-third of the stocks of local mature forests. Therefore, the carbon sink potential of these forests is not exhausted yet. In general, species richness, structural diversity, and community-weighted mean wood density had significant effects on annual rates of AGB gain and loss in the ERF and AF, presumably mediated by forest development stage. Greater gain rates were found on plots with smaller structural diversity and wood density in both forest types. Likewise, greater loss rates were found on plots with smaller structural diversity and wood density in the ERF, suggesting that light-demanding species are being substituted by more selective species. Conversely, greater loss rates were not found on plots with smaller structural diversity and wood density in the AF, possibly because light-demanding species are still finding favorable environments for their persistence. Our findings can contribute to reducing uncertainties in large-area estimates of AGB stocks, growth, and carbon uptake rates. The implementation of large-scale inventories should be encouraged to improve carbon cycle models under different climate change scenarios.